A carbon arc furnace for silicon needs at least 10 kW of electrical power, which at the 40 to 80 V of a carbon arc is a current of 125 to 250 A, so the supply, the conductors and the arc control are a larger project than the furnace body itself. No source gives furnace dimensions, so the guidance here is limited to what the sources do support: the power arithmetic, the electrodes, the arc control and the ventilation.

The electrical and mechanical requirements of a small carbon arc furnace for reducing quartz to silicon.

The Power Supply Is The Hard Part

The carrier says a small arc furnace needs “at least 10 kW” and calls this a chicken-and-egg problem, since electricity is needed to make solar panels; its answer is to start with hydroelectric or wind power. It also asks for a furnace reaching 1,800–2,000°C, while Silicon Purification says 1,400–1,700°C (Quartz Reduction notes the disagreement). The only electrical figures for a carbon arc in the corpus are for lamps: 40–80 V DC at 5–25 A (Carbon Arc Lamp). A furnace arc is not a lamp arc and the corpus gives no furnace arc voltage, so the lamp voltage range is a stand-in and should be replaced by measurement.

An arc has negative resistance: more current lowers the arc voltage, so without a series ballast the current runs away (Ballast Function). A resistor ballast works on DC but wastes power.

What You Need

Method: Plan The Furnace

Step 1: Fix the power and voltage. Choose the arc voltage by measuring a trial arc, then compute the current from P = V × I.

Step 2: Size the conductors. The wire-selection table stops at 1/0 AWG (150 A in free air), so a furnace current above that needs 2/0 or larger, several cables in parallel, or copper bar at 1.5–2 A per mm² (Bus Bars).

Step 3: Choose a ballast or a controlled supply. Compute the loss as below and prefer a supply whose voltage is near the arc voltage.

Step 4: Arrange feed and regulation. Electrodes burn away and the gap widens; a counterweighted gravity feed held by a clutch in the arc circuit is the hand-built option in Arc Regulation.

Step 5: Provide ventilation and a hood for carbon monoxide and arc light, before the first heat.

Step 6: Test cold, with a small trial charge at a low current before any full run.

Worked Example: Current, Conductors And Ballast

Current. I = P ÷ V for 10,000 W: at 80 V, 125 A; at 60 V, 167 A; at 40 V, 250 A.

Conductors. At 167 A the bus-bar rule gives a copper cross-section of 167 ÷ 2 = 84 mm² to 167 ÷ 1.5 = 111 mm². A bar of 25 mm × 6 mm is 150 mm², comfortably over that. The wire-selection table’s largest free-air entry, 1/0, is 150 A, below 167 A.

Ballast loss. If the supply is 100 V and the arc runs at 60 V and 167 A, the resistor must drop 40 V: R = 40 ÷ 167 = 0.24 Ω, dissipating 40 × 167 = 6,680 W. That is 6.7 kW burnt as heat in the ballast on top of the 10 kW delivered to the arc. A supply near 65 V would drop only 5 V across a ballast: 5 × 167 = 835 W.

Supply scale. The wind-power article gives about 345 W from a 2 m rotor at 8 m/s. Then 10,000 ÷ 345 = 29 such rotors, so wind alone is not realistic; a hydro site is the likelier source.

Common Mistakes

  • Using lamp-grade wire — at 167 A thin cable heats and burns
  • No ballast or regulation — the current runs away when the arc steadies
  • Wasting power in a resistor — match the supply voltage to the arc instead
  • Letting the electrode gap grow unattended — the arc dies, then restarts as a short

Safety

A 10 kW arc is a high-current burn, shock and fire hazard

At 60 V and 167 A, a loose joint carries enough current to melt a connector, and the lethal-current article puts 50 to 120 V in the potentially serious range for wet skin. The arc emits ultraviolet that causes arc eye (Carbon Arc Lamp), molten silicon sits above 1,414°C, and the reaction releases carbon monoxide. Enclose the live conductors, insulate the electrode holders, keep the hood or open air working before any run, wear welding-grade eye protection, and never work on the circuit with the supply connected (Lethal Current).

What’s Next